CLF-C02 Cloud Technology and Services Practice Question

Answer analysis

Option-by-option breakdown

For each option: why learners choose it and why it is or isn't the right answer here.

• ✗

  Amazon Textract

  Why it's wrong here

  Amazon Textract is an OCR and document analysis service that extracts text, forms, and tables from scanned documents, including handwriting. However, its output is unstructured or semi-structured text with no domain-specific understanding of medical concepts. It cannot interpret clinical relationships such as drug-to-dosage associations, nor does it detect PHI in a medical context, so it is not suitable for deriving meaningful clinical insights from patient records.

• ✗

  Amazon Comprehend

  Why it's wrong here

  Amazon Comprehend is a general-purpose natural language processing (NLP) service that identifies sentiment, key phrases, entities (e.g., people, places), and language in text. Its generic entity recognizer treats medical terms as ordinary words and lacks the clinical ontology needed to distinguish a diagnosis from a symptom or to recognize medication dosages and administration routes. Because it was not trained on medical corpora with PHI-specific annotations, it would miss or misclassify the domain-specific entities that a healthcare workflow requires.

• ✓

  Amazon Comprehend Medical

  Why this is correct

  Amazon Comprehend Medical is a specialized NLP service built with pre-trained models on medical and clinical text. It extracts entities such as medications, conditions, procedures, and dosages, and it also identifies protected health information (PHI) like patient names and addresses. It understands medical relationships—for example, linking a drug name to its dosage and route—making it the correct choice for processing clinical notes. Unlike generic services, it is purpose-built for healthcare use cases such as medical coding, clinical trial matching, and PHI redaction.

• ✗

  Amazon Rekognition

  Why it's wrong here

  Amazon Rekognition is a computer vision service that performs facial analysis, object and scene detection, and content moderation on images and videos. Its pre-trained models are designed for visual features, not for processing medical text or extracting clinical entities. Therefore, applying Rekognition to clinical notes or scanned medical documents would fail to identify medications, conditions, or protected health information (PHI).

Go deeper

Related to this question